Don't let Google know I'm lonely!

TL;DR

This paper introduces a new privacy metric called {}-indistinguishability to evaluate how well online systems can detect sensitive topic profiling, demonstrating high detection rates with practical tools.

Contribution

It proposes a scalable, practical method for assessing online privacy disclosure and profiling risks based on a novel -indistinguishability framework.

Findings

Achieved over 98% detection rate for sensitive topics

Developed scalable tools for privacy assessment

Validated results using publicly available resources

Abstract

From buying books to finding the perfect partner, we share our most intimate wants and needs with our favourite online systems. But how far should we accept promises of privacy in the face of personal profiling? In particular we ask how can we improve detection of sensitive topic profiling by online systems? We propose a definition of privacy disclosure we call {\epsilon}-indistinguishability from which we construct scalable, practical tools to assess an adversaries learning potential. We demonstrate our results using openly available resources, detecting a learning rate in excess of 98% for a range of sensitive topics during our experiments.